> For the complete documentation index, see [llms.txt](https://comunity.gitbook.io/learning.comunitynexus/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://comunity.gitbook.io/learning.comunitynexus/26.2/toolkit-guides/intelligence/connect-agents.md).

# Connect agents

An agent can call another agent. There are two mechanisms, and they differ in what they reach.

{% hint style="info" %}
There is no Connect agents entry in the Intelligence menu. This page covers the choice between the two mechanisms. Agent handoff is set up under **Intelligence (Preview)**, then **Triggers**, as **Agent Handoff Rules**; agent to agent carries its own connection and authentication.
{% endhint %}

## Before you start

* At least two agents, where both are in this project and you intend to use a handoff. See Create an agent.
* The connection and authentication details for the other agent, where it is outside this project.

## Agent handoff

A handoff links a source agent to a secondary agent, so the first can call the second. This is how agents are composed within a project. Both are already defined in the Toolkit, so nothing beyond the link itself has to be configured.

Handoff rules are defined under **Triggers**, as **Agent Handoff Rules**, which set the transition rules between agents.

Use a handoff where a coordinating agent needs specialists. An agent receiving a fault hands off to one agent to classify it and another to decide whether it escalates, and each of those is instructed and tested on its own.

## Agent to agent

Agent to agent is Azure's own mechanism for agents communicating, and it carries its own connection and authentication configuration.

Its use is reaching an agent that is not yours. Where a service provider publishes a fraud agent, agent to agent is how your agent calls theirs, and your agent becomes the coordinator across both.

{% hint style="info" %}
Within a project, a handoff does the same work with considerably less configuration. Reserve agent to agent for agents outside the project.
{% endhint %}

## The A2A protocol description

The A2A description states what your agent can do for other agents, as distinct from what it does.

The two are not the same thing. Instructions tell your agent how to carry out its role. The A2A description tells another agent, which knows nothing about your project, what it can ask of yours and what it will get back in return. An agent that is never called by another agent does not need one.

## Where to go next

* Trigger an agent: start the chain from a schedule or a data change
* Add knowledge to an agent: give a coordinating agent the policy it applies